Symbiotic Planning for Planetary Exploration

Abstract

Planetary exploration missions avoid the destinations that offer the greatest scientific payout because they entail risks too great for a primary rover. Given the high costs of sending to extraterrestrial bodies, even the small possibility of losing a primary rover precludes the exploration of important, yet risky features. The solution to this problem is a new paradigm for planetary exploration wherein the primary rover is accompanied by multiple cheaper, lighter, and more expendable companion rovers with differing capabilities. These rovers would complement each other’s strengths and weaknesses through a strategy called Symbiotic Exploration. Several algorithmic challenges must be solved before symbiotic exploration in planetary environments can become a reality. The less featured companion rovers will often lack a direct line of communication to Earth, requiring that they remain within communication distance of the lander or more capable primary rover to relay data. Similarly, the secondary rovers may sacrifice the large, expensive solar panels and electric generators for a high-density rechargeable battery. Such a configuration would require resource aware planning, as well as occasional rendezvous with a rover or base station capable of recharging or swapping out that battery. A symbiotic planner must be able to develop plans in highly dynamic environments while meeting these constraints. Distributed Path Consensus (DPC) is an iterative approach to enforcing time-dependent multi-rover rendezvous constraints. It is limited, however, in that it can only apply such constraints based upon the time domain, and the value at which those constraints apply must be identical between the two rovers. This research proposes an evolution of the DPC algorithm (called DPC.TF) that allows it to enforce rendezvous constraints based upon any arbitrary resource, including those that are non-monotonic, even when their values differ across rovers. Furthermore, it allows for complex rendezvous and maximum separation distance constraints to be specified through an intuitive syntax. A planner based around the DPC.TF algorithm was developed, and its correctness proved on synthetic data. The planner generated routes using actual Lunar data for a variety of multiple rover configurations and symbiotic constraints. This research was able to show that routes do exist to high-interest, permanently shadowed sites while maintaining symbiotic constraints. Furthermore, capabilities that would be required of each rover to explore these sites was analyzed and determined. Such regions have been previously considered inaccessible but, through the paradigm of Symbiotic Exploration, can be thoroughly explored.